¹ In this paper I use the system of transliteration employed in my Hittite Glossary; that is, I give a minimum of phonetic interpretation, but indicate the orthography of the monuments. In this system a macron over a vowel indicates that the vowel sign is inserted alongside a syllabic sign; it does not necessarily mean anything for the quantity of the vowel. In case it seems important to indicate the pronunciation of a word more precisely, I enclose a phonetic transcription in square brackets. The reasons for the phonetic interpretations thus given may be found in my Comparative Grammar of the Hittite Language.

^1a So also bu-iš-nu-uš-ki-e-it (KBo. 3.28.2.18).

² I know full well that analogical sources can be dug up for most inconvenient forms, given sufficient energy and ingenuity; but I think that it will be difficult to dispose of all these words by that means.

³ Cf. Paul, Prinzipien der Sprachgeschichte⁵ 67.

⁴ JAOS 44.38-53.

⁵ It is no part of our present purpose to treat vowel assimilation in Greek. We merely take the instances listed by Brugmann as typical. For other examples, see J. Schmidt, KZ 32.321-94; Ehrlich, Untersuchungen über die Natur der griechischen Betonung 127-31; Walde, Streitberg Festschrift 152-70.

⁶ See Sturtevant, CP 8.337-9.

⁷ I plan to treat elsewhere the peculiar liability to assimilation and dissimilation of reduced vowels (IE and ) in Greek.

⁸ The only way to avoid this dilemma, I think, would be to point out a clear instance of an assimilative or dissimilative change as frequent as the assumed Hittite change, which should interfere with morphological systems without applying regularly to isolated forms.

⁹ Lang. 3.161-8.

¹⁰ RHA 1.76-88.

¹¹ See especially pp. 92, 97-9, 220, 221 f., 234 f., 237, 239-41.

¹² JAOS 54.167 fn. 9; Lang. 10.206.

¹³ Benveniste, BSL 33.142 fn.1, and Meillet, BSL 34.30, object to the connection of eku- ‘drink’ with Lat. aqua and ON gir ‘sea’ on the ground that words for ‘water’ and for ‘drink’ are never related. I doubt whether there is any difficulty in a direct connection between the meanings, but it is perhaps worth while to point out a possible indirect connection. The Hittite verb means also ‘cause to drink, provide with drink’, and if this is the primary meaning we have a close parallel in English to water (animals). To be sure I cannot cite Hittite aku-with a word denoting an animal as object. Hrozný, AOr. 3.447, translates KUB 1.13.2.30: na-at ar-ha a-ku-wa-an-zi, ‘Et ils (le) leur < i.e. the horses > font boire entièrement’; but the sentence pretty certainly means ‘and they (the horses) drink (it) all up’. This passage does show, however, that the verb could be used of animals, and there are many passages where it means ‘give drink to (a person)‘.

¹⁴ See HG 258-60.

¹⁵ Edgerton, Lang. 10.235-65, has undertaken to prove that schwa secundum never existed. To answer his argument will require another paper; but in the meantime I must say that Edgerton has convinced me that we must not distinguish a reduced grade from a zero grade in the short-vowel series. On the contrary we have two forms of the reduced grade (zero or ) according to the nature of the surrounding sounds. When unaccented short vowels were lost in all convenient positions, a minimal vowel (b) remained wherever its loss would have produced consonant plus consonant or consonant plus semivowel, either initially or after a long syllable. This rule gives us, for the first time, a workable distinction between (e.g.) pd and pd, tn and t .

¹⁶ Le cinquième déclinaison latine = Det Kgl. Danske Videnskabernes Seiskab, Historisk-fìlologiske Meddelelsen 11.5, especially 22-6.

¹⁷ Petersen, JAOS 54.168.